Classification To Predict Credit Card Application Acceptance Using Support Vector Machine
Asfan Muqtadir (*), Andik Adi Suryanto, Amaludin Arifia, Fitroh Amaluddin

Faculty of Enginering, Department of Informatics Engineering, PGRI Ronggolawe University

Abstract

Credit card creation is one of the banking services that has a large source of risk for business operations. The process of granting credit card functions includes application and customer profile analysis. Customer profile analysis can be determined based on many factors such as savings owned, transfers or cash flow from customer accounts, and income from customer credit applications. This is done as an implementation of the SVM algorithm for the classification of credit card application acceptance using data taken from the Kaggle website. In this research, the SVM method is used with an additional function, namely a kernel trick. From the evaluation of the classification model along with the four kernel functions using the confusion matrix, it is found that the sigmoid kernel has the highest precision and recall percentage of 0.982 and recall 0.985, while the highest accuracy is produced by the Polynomial Kernel of 98%.

Keywords: classification- Support Vector Machine- kernel trick- confusion matrix

Topic: Engineering